编译原理之词法分析和语法分析

最新推荐文章于 2022-03-04 22:59:39 发布

刘秋杉

最新推荐文章于 2022-03-04 22:59:39 发布

阅读量5.7k

点赞数 2

分类专栏： C 文章标签：编译原理递归下降实验代码词法分析语法分析

本文链接：https://blog.csdn.net/BlueCloudMatrix/article/details/41701233

版权

本文介绍了使用有限自动机进行词法分析和通过递归下降法进行语法分析的过程，结合EBNF文法生成中间代码。通过实际编写代码，有助于深入理解编译原理的基本概念。

摘要由CSDN通过智能技术生成

花了一天写出的程序没有顾及很多层面，但对于理解基本的实验道理和交上实验还是有点帮助的。代码实现了基于有限自动机的词法分析，采用递归下降分析法和EBNF文法实现语法分析并生成中间代码。from sdu

lexAnalysis.h

/*
 * lexAnalysis.h
 *
 *  Created on: 2014-12-2
 *      Author: liuqiushan
 */

#ifndef LEXANALYSIS_H_
#define LEXANALYSIS_H_


#include <stdio.h>
#include <stdlib.h>
#include <string.h>

typedef enum _SYM
{
		_CONST,
		_VAR,
		_procedure,
		_begin,
		_end,
		_if,
		_then,
		_ID, // such as the 'a' in 'var a;'
		_INT, // such as the '10' in 'const a = 10;'
		_ASSIGN,// '='
		_PLUS,// '+'
		_SUB,// '-'
		_STAR,// '*'
		_DIV,// '/'
		_LESS,// '<'
		_MORE,// '>'
		_LESSEQ,// '<='
	    _MOREEQ,// '>='
	    _DH,// ','
	    _MD,// ':='
	    _LEFT,// '('
	    _RIGHT,// ')'
	    _JH,// '#'
	    _FH// ‘;’
}SYM;

SYM getSYM(char* str);

void appendStr(char* str, char* c, int p_str);

int isABC(char* c);

int isNumber(char* c);

#endif /* LEXANALYSIS_H_ */

LexAnalysis.c

/*
 * LexAnalysis.c
 *
 *  Created on: 2014-12-2
 *      Author: liuqiushan
 */

#include "lexAnalysis.h"

SYM getSYM(char* _str)
{
	SYM temp = _ID; // default identifier
	if(!strcasecmp(_str, "CONST"))
		temp = _CONST;
	else if(!strcasecmp(_str, "VAR"))
		temp = _VAR;
	else if(!strcasecmp(_str, "procedure"))
			temp = _procedure;
	else if(!strcasecmp(_str, "begin"))
				temp = _begin;
	else if(!strcasecmp(_str, "end"))
				temp = _end;
	else if(!strcasecmp(_str, "if"))
					temp = _if;
	else if(!strcasecmp(_str, "then"))
					temp = _then;

	return temp;
}

void appendStr(char str[], char* c, int p_str )
{
	str[p_str] = *c;
}

int isABC(char* c)
{
	char ch = *c;

	if(ch=='q'||ch=='w'||ch=='e'||ch=='r'||ch=='t'||ch=='y'||ch=='u'||ch=='i'||ch=='o'||ch=='p'||
					ch=='a'||ch=='s'||ch=='d'||ch=='f'||ch=='g'||ch=='h'||ch=='j'||ch=='k'||ch=='l'||
					ch=='z'||ch=='x'||ch=='c'||ch=='v'||ch=='b'||ch=='n'||ch=='m'||ch=='Q'||ch=='W'||
					ch=='E'||ch=='R'||ch=='T'||ch=='Y'||ch=='U'||ch=='I'||ch=='O'||ch=='P'||ch=='A'||
					ch=='S'||ch=='D'||ch=='F'||ch=='G'||ch=='H'||ch=='J'||ch=='K'||ch=='L'||ch=='Z'||
					ch=='X'||ch=='C'||ch=='V'||ch=='B'||ch=='N'||ch=='M')
				return 1;
			else
				return 0;
}

int isNumber(char* c)
{
	char ch = *c;

	if(ch=='1'||ch=='2'||ch=='3'||ch=='4'||ch=='5'||ch=='6'||ch=='7'||ch=='8'||ch=='9'||ch=='0')
				return 1;
			else
				return 0;
}

synAnalysis.h

/*
 * synAnalysis.h
 *
 *  Created on: 2014-12-2
 *      Author: liuqiushan
 */

#ifndef SYNANALYSIS_H_
#define SYNANALYSIS_H_

typedef enum _KIND
{
	PRO,//procedure
	CON,//constant
	VAR //variable
}CVPKind;

/*
 * PL/0 Object Instruction
 *
 * There are three domains:
 *
 * f is the function code
 * l is the difference between the declared level and the invoked level of the variable
 *
 * -----------------------------
 * |           |       |       |
 * |     f     |   l   |   a   |
 * |           |       |       |
 * -----------------------------
 */
typedef enum _FunctionCode
{
	LIT,//① LIT: put the constant on the top of the stack, a is the constant
	LOD,//② LOD: put the variable on the top of the stack,